We recently had a critical app go down for almost two full days. Work stopped, the team built workarounds in spreadsheets, and someone asked for an update every few hours.
Outages happen. I’ve been on the vendor side of this. Every B2B company that runs long enough will have a bad day, and some will have a bad two days. I can forgive the downtime.
What I can’t get past is that they never offered a service-level credit. Not during the outage, not after, not in the postmortem email.
It’s so little money, and they never offered. I won’t forget that.
Two Days of Credit on a $60K Contract Is About $330
Run the math on a typical mid-market contract. A customer paying $60,000 a year is paying about $165 a day. Two days down is roughly $330. Credit a full week and it’s about $1,150. Credit a full month and it’s $5,000, against a $60,000 renewal.
For most vendors, that’s less than one field marketing dinner, and less than the CSM hours that will go into defending the account at renewal.
I don’t need $330. Nobody at our company needs $330. The credit tells the customer you know it hurt them and you took responsibility without being asked.
When the Customer Has to Ask, the Outage Becomes a Relationship Problem
One commenter on my LinkedIn post said making the customer ask for compensation turns it into a relationship failure. I agree.
The outage is an engineering failure, and customers understand those. Every one of us has shipped a bad deploy.
After the outage, though, the customer is stuck. They can dig up the SLA, file a ticket, or email their rep and feel petty doing it. Most won’t do any of that. They’ll quietly note it.
That note comes back 9 months later, when someone on the team suggests looking at alternatives and nobody pushes back.
Same dollar amount, very different message. If the vendor offers it first, the customer sees that someone was paying attention. If the customer has to request it, they see that nobody was. If it never comes, they assume nobody cared.
If the Credit Didn’t Fire Automatically, Nobody Owns Uptime
Another commenter made a point I think most vendors miss: the SLA credit shows the customer you actually measure your own reliability.
If you have an SLA, you had a two-day outage, and no credit went out, then one of three things is true:
- You don’t track uptime against the SLA
- You track it, but nobody is responsible for acting on it
- Someone is responsible, and they decided not to bother
The second is the most common, and you can fix it in an afternoon.
For systems a business runs on (CRM, billing, ERP), buyers ask about this in contract review. Who declares an outage? How long before credits kick in? Are they automatic, or does the customer have to claim them? Where is it written down? Vague answers cost you before you ever go down.
Your Champion Has to Defend You at Renewal
I sign a lot of vendor contracts across SaaStr and the portfolio, and this is how it goes on the customer side.
The outage happens and the team is annoyed. Then things go back to normal, and within a few weeks nobody thinks about it day to day.
The champion remembers, though. When renewal comes up, or a competitor’s AE sends a well-timed email, or a new exec asks why we use this tool, the champion doesn’t say “they had an outage.” They say “they had a two-day outage and didn’t offer us anything.”
That’s much harder to defend internally, and it makes your champion look like they picked a vendor that doesn’t respect them.
Vendors who skip a $330 credit often spend 10x to 100x that later on save campaigns, renewal discounts, and executive escalations.
What We’d Want to See After a Major Outage
If you run a B2B or B2B + AI company, here’s what to do. None of it is expensive.
- Don’t make the customer ask. If a customer has to file a ticket or email their rep for a credit, you’ve put the burden on the person who already got hurt. They have to find the SLA, do the math, and feel petty asking. Most won’t, and they’ll remember. Calculate the credit yourself, apply it, and tell them it’s done. If the SLA math produces a tiny number, round up generously.
- Do it within 24 hours, with an honest root cause analysis. Don’t wait for the next invoice cycle 60 days out. A credit that shows up two months later with no note attached barely registers. Send the credit and the root cause in the same email, within a day of the outage being resolved. “We experienced an infrastructure issue” tells the customer nothing. Say what broke, why, and what you changed so it doesn’t happen again.
- Thank them and say you’ll do better. Thank them for their patience and for staying with you. Then say plainly that you’ll do better. Skip “we apologize for any inconvenience this may have caused.” For a multi-day outage on a critical system, this should come from a VP or the CEO, not a no-reply address.
Then automate it. If you have an SLA, connect uptime monitoring to billing so credits go out automatically when a threshold is crossed. Nobody on your side has to remember, and no customer has to ask.
For a typical customer, all of this costs a few hundred dollars and 30 minutes of an executive’s time.
What That Vendor Could Have Done in One Email
Two days down is bad. Two hours is rought, but two days? But anyhow, things break, and I’ve been there. It was a core workflow they broke in a patch, not the entire app. It wasn’t down for everyone.
One email within 24 hours would have done it: a credit, an honest explanation of what went wrong, and a thank you for sticking with them. It would have cost less than a nice dinner. Instead, the missing credit is the first thing I think of when their name comes up.
If your app goes down, send that email before anyone asks.
